Intensive Care Med (2003) 29:10161019 DOI 10.

1007/s00134-003-1719-x

B R I E F R E P O RT

Roberto Imberti Guido Bellinzona Francesca Riccardi Michele Pagani Martin Langer

Cerebral perfusion pressure and cerebral tissue oxygen tension in a patient during cardiopulmonary resuscitation

Received: 29 October 2002 Accepted: 18 February 2003 Published online: 28 March 2003 Springer-Verlag 2003

R. Imberti () G. Bellinzona F. Riccardi M. Pagani M. Langer Servizio di Anestesia e Rianimazione II, IRCCS Policlinico S. Matteo, 27100 Pavia, Italy e-mail: r.imberti@smatteo.pv.it Tel.: +39-0382-502769 Fax: +39-0382-526383

Abstract Objective: To report on the effects of cardiopulmonary resuscitation (CPR) instituted immediately after a cardiac arrest on cerebral perfusion pressure (CPP) and cerebral tissue oxygen tension (PbrO2). Design: Case report. Setting: ICU of a university hospital. Patient: A head-injured 17-yearold man submitted to multimodal neurological monitoring underwent sudden cardiac arrest and successful CPR. Interventions: External chest compression, 100% oxygen ventilation, volume expansion and standard ACLS protocols. Measurements and results: Heart rate, ECG, mean arterial blood pressure (MABP), ETCO2, PaO2, intracranial pressure (ICP), CPP and PbrO2 were continuously monitored during CPR and data recorded at 15-s intervals by a dedicated personal computer. At the onset of the cardiac arrest, PbrO2 decreased to zero. The institution of CPR resulted in a progressive increase of MABP, CPP and PbrO2. Assuming, on the basis of previous experimental and clinical reports, 8 mmHg PbrO2 as a possible isch-

aemic/hypoxic threshold value, during the first 6.5 min of CPR, PbrO2 values were below this threshold (range 07 mmHg) and CPP values were <25 mmHg for 81.5% of the time. In the following 5.5 min, more efficient CPR generated CPP values >25 mmHg for 77.3% of the time. These values were associated with a PbrO2 >8 mmHg (range 828 mmHg) at all times. Conclusions: In the clinical setting of a witnessed cardiac arrest, immediate institution of CPR can be effective in generating PbrO2 values above a supposed ischaemic/hypoxic threshold when CPP is >25 mmHg. PbrO2 monitoring by the Licox system is sensitive and reliable, even at low values, and can be suitable for evaluating cerebral oxygenation during experimental CPR. Keywords Cardiac arrest Cardiopulmonary resuscitation Brain tissue partial pressure of oxygen Intracranial pressure Cerebral perfusion pressure Cerebral oxygen metabolism

Introduction
Prevention of cerebral damage by maintaining and restoring adequate oxygenation is one of the goals of cardiopulmonary resuscitation (CPR) [1, 2]. Attempts to evaluate cerebral oxygenation by measuring cerebral tissue oxygen tension (PbrO2) during CPR have been un-

successful in animals [3] and, to the best of our knowledge, PbrO2 changes during CPR have never been documented in humans. We report on a patient affected by severe traumatic brain injury who suffered a cardiac arrest while staying in our ICU. In this patient, whose PbrO2 was being recorded as part of a multimodal monitoring process, CPR

1017

instituted immediately after the cardiac arrest was effective in achieving PbrO2 values >8 mmHg when CPP values were >25 mmHg. Based on previous experimental and clinical reports [4, 5, 6, 7, 8], 8 mmHg PbrO2 was adopted as a possible ischaemic/hypoxic threshold value.

Case report
A 17-year-old man with multiple injuries was admitted to our hospital after being involved in a car accident. On admission the patient was unconscious. The brain CT scan showed an acute left temporo-parietal subdural haematoma, midline shift, right frontal and temporal contusions and effacement of the cortical sulci. The subdural haematoma was promptly evacuated. Patient management also included mechanical ventilation, sedation, mannitol and NaCl 7.5% infusion, control of volaemia and ABP, targeted on intracranial pressure (ICP), cerebral perfusion pressure (CPP) and SjvO2 and PbrO2 values. A CT scan performed approximately 24 h after brain injury showed the appearance of a diffuse hypodensity in the left hemisphere. Heart rate (HR), mean arterial blood pressure (MABP), ICP, CPP, end-tidal CO2 (ETCO2), arterial PO2 (PaO2) and PbrO2 were continuously monitored between day 1 and day 10 after injury. The MABP transducer was positioned at the level of the ear. The ICP was monitored by an intraparenchymal fiberoptic catheter (Camino Laboratories, San Diego, Calif.) placed in the right frontal lobe. PbrO2 was measured by a miniaturized Clarktype electrode (Licox, GMS, Kiel, Germany) inserted in an uninjured area of the cerebral white matter, 11.5 cm behind the ICP catheter, as confirmed by CT scan. A second Clark-type electrode (Licox, GMS, Kiel, Germany) was inserted through an 18-G cannula into a femoral artery for continuous monitoring of PaO2. Each probe was calibrated according to the manufacturers specifications. All parameters were recorded by a multiparametric monitor (Merlino, Hewlett Packard, Bblingen, Germany) and transmitted as analog signal to a dedicated personal computer using an adapted software (LabView, National Instruments, Austin, Texas), which recorded all the data at 15-s intervals. The CPP was calculated on-line (CPP=MABPICP). SjvO2 was monitored by intermittent blood sampling. On the ninth day, the patient, whose Glasgow Coma Scale (GCS) score was 5 (m3, v1, e1), had an iatrogenic cardiac arrest with asystole due to unintentional rapid infusion of a concentrated solution of potassium instead of hypertonic saline (NaCl 7.5%) solution. One minute before the cardiac arrest, the patients MABP was 94, ICP 36, CPP 58 and PbrO2 15 mmHg. The CPR was started after a few seconds and FiO2 was switched to 1.0. Plasma expanders, epinephrine, calcium chloride and sodium bicarbonate were promptly administered through a central line. The onset of cardiac arrest resulted in a steep reduction of CPP which dropped to 0 mmHg. PbrO2 decreased from 15 to 0 mmHg in 2 min, and ICP decreased from 35 to 15 mmHg (Fig. 1). The institution of CPR resulted in a progressive increase in MABP, CPP and PbrO2. After restoration of spontaneous circulation, which occurred after 12 min, MABP, ICP and CPP increased abruptly, peaking at 180, 60 and 120 mmHg, respectively, before gradually returning towards baseline in the following minutes. The increase of MABP and CPP, while the patient was still being ventilated with 100% oxygen, was associated with a steep increase in PaO2 (peak 440 mmHg) and a parallel increase in PbrO2 (peak 185 mmHg; Fig. 1). During the cardiac arrest, ETCO2 behaviour was consistent with the circulatory effects of CPR (Fig. 1). Forty-eight hours after this event, the patients GCS score and EEG were unchanged from those immediately preceding the cardiac arrest, as was a cerebral CT scan performed 2 days later; how-

Fig. 1 Heart rate, ETCO2, mean arterial blood pressure (MABP), intracranial pressure (ICP), cerebral perfusion pressure (CPP), PaO2 and cerebral tissue oxygen tension (PbrO2) before cardiac arrest, during 12 min of cardiopulmonary resuscitation (CPR) and after restoration of spontaneous circulation. The gray strip in the lower part of the figure indicates PbrO2 values <8 mmHg ever, energetic and functional derangement and subtle anatomical cerebral damage cannot be excluded. At a follow-up performed 2 years later, the patient was hemiplegic, conscious, and could talk and could eat autonomously.

Discussion
Direct measurement of cerebral tissue oxygen tension (PbrO2) has gained popularity as part of the monitoring of patients with severe traumatic and haemorrhagic brain injuries [6, 7, 8, 9, 10]. PbrO2 monitoring can be performed during a cardiac arrest and CPR only in experimental studies or occasionally, as in our experience, in patients suffering a cardiac arrest while already undergoing multimodal neurological monitoring because of severe brain injury. The on-line registration of all parame-

1018

Fig. 2 Relationship between CPP and PbrO2 during cardiac arrest and CPR. The lower dotted line at 8 mmHg PbrO2 indicates the possible ischaemic/hypoxic threshold. The upper dotted line is positioned at 25 mmHg CPP

Fig. 3 Relationship between CPP and PbrO2 during cardiac arrest and CPR. Values are plotted adopting a time delay of 1.7 min between CPP and PbrO2 changes (see Discussion)

ters (HR, MABP, ICP, CPP, PbrO2) before the asystolic cardiac arrest and during CPR until restoration of spontaneous circulation (12 min) gave us the unique opportunity to analyze the effects of CPR not only on haemodynamics, but also on PbrO2. Animal studies have attempted to investigate this issue [3], but the experimental setup did not allow very low values of PbrO2 to be measured; however, in our experience and that reported by others, the Licox system of monitoring cerebral tissue oxygenation is able to measure even very low PbrO2 values [6, 8, 9, 10]. We found [9] that after 49 days of PbrO2 monitoring, the zero drift was 0.51.5 mmHg and the sensitivity drift was 4.53.7, indicating that measurements are reliable. The effects of CPR on haemodynamics and cerebral oxygenation are well illustrated in records reproduced in Fig. 1. There is an expected relationship between the mean arterial blood pressure generated by external chest compression, volume expansion and repeated epinephrine administration, and CPP and expired CO2. The PbrO2 tracing of particular interest is: increased PaO2 has a very slight effect, but the CPP has a clear, albeit slightly delayed, effect on PbrO2. The relationship between CPP and PbrO2 during CPR is better detailed in Fig. 2. Changes in cerebral perfusion generally precede changes in oxygenation. In our patient, the fall of CPP due to cardiac arrest is followed, after 2.0 min, by a zero value of PbrO2. The highest level of CPP during CPR, reached at minute 8, is followed by the highest level of PbrO2 1.5 min later. If we accept a time delay of 1.7 min between CPP and PbrO2 values (1.7 being an intermediate value between 2.0 and 1.5 min), in our patient a CPP value of 25 mmHg led to a PbrO2 value of 8 mmHg.

Again, adopting 1.7 min as time delay between CPP and PbrO2 changes during CPR, a strong correlation was found between CPP and PbrO2 (r2=0.719; Fig. 3). A similar delay has been reported in a recent paper by Lang et al. [11]. From our specific and single case we cannot answer the most relevant question: For what proportion of time during CPR was oxygen delivery adequate? Experimental studies in animals performed under conditions of cerebral hypoxia and ischaemia [4, 5] and clinical studies performed in patients with severe traumatic brain injury using the Licox system [6, 8] indicate that the risk of cerebral ischaemic damage is very high when the PbrO2 is below 810 mmHg. In another experimental investigation on CPR in dogs [11], a CPP value of 25 mmHg was shown to maintain supratentorial blood flow and ATP levels at 6070% of normal. If we therefore accept a hypothesis of minimal levels of adequacy as being 25 mmHg for CPP [12] and 8 mmHg for PbrO2 [4, 5, 6, 8], during the first 6.5 min of CPR, CPP values were <25 mmHg for 81.5% of the time (average over the period 15.310.5 mmHg, and PbrO2 values were always <8 mmHg). In the following 5.5 min, with improving CPR and generation of higher MABP values, PbrO2 values were always >8 mmHg (range 828 mmHg) and CPP, despite a slight increase in ICP, was >25 mmHg for 77.3% of the time, averaging 308 mmHg. Although no conclusion on adequacy of oxygen supply can be drawn in our patient, the two hypothesized threshold values (PbrO2=8 mmHg and CPP=25 mmHg) seem to be closely linked and are consistent with excellent experimental and clinical studies [4, 5, 6, 8, 12].

1019

Conclusion
In conclusion, our case report shows, for the first time, that during CPR there is a close relationship between CPP and PbrO2. Although a single case can add limited information on the supposed adequacy of cerebral oxygenation/perfusion when PbrO2 values are >8 mmHg and

CPP values are >25 mmHg, our findings, together with those of the above-mentioned experimental and clinical studies, suggest that these levels could be useful for setting haemodynamic goals that should be reached during CPR.
Acknowledgements Acknowledgement. This work was supported by a grant from IRCCS Policlinico S. Matteo, Pavia, Italy.

References
1. Safar P (1986) Cerebral resuscitation after cardiac arrest: a review. Circulation 74 (Suppl IV):138153 2. International Guidelines 2000 for CPR and ECC (2000) A consensus on science. Resuscitation 46:1448 3. Nozari A, Rubertsson S, Gedeborg R, Nordgren A, Wiklund L (1999) Maximization of cerebral blood flow during experimental cardiopulmonary resuscitation does not ameliorate post-resuscitation hypoperfusion. Resuscitation 40:2735 4. Farrar JK (1991) Tissue PO2 threshold of ischemic cell damage following MCA occlusion in cats. J Cereb Blood Flow Metab 11 (Suppl 2):S553 5. Kariman K, Hempel FG, Jobsis FF (1983) In vivo comparison of cytochrome a/a3 redox state and tissue PO2 in transient anoxia. J Appl Physiol 55:10571063 6. Kiening KL, Unterberg AW, Bardt TF, Schneider G-H, Lanksch WR (1996) Monitoring of cerebral oxygenation in patients with severe head injuries: brain tissue PO2 versus jugular vein oxygen saturation. J Neurosurg 85:751757 7. Dings J, Meixensberger J, Amschler J, Hamelbeck B, Roosen K (1996) Brain tissue PO2 in relation to cerebral perfusion pressure, TCD findings and TCDCO2 reactivity after severe head injury. Acta Neurochir 138:425434 8. Gopinath SP, Valadka AB, Uzura M, Robertson CS (1999) Comparison of jugular venous oxygen saturation and brain tissue PO2 as monitors of cerebral ischemia. Crit Care Med 27:23372345 9. Imberti R, Bellinzona G, Langer M (2002) Cerebral tissue PO2 and SjvO2 changes during moderate hyperventilation in patients with severe traumatic brain injury. J Neurosurg 96:97102 10. Bruzzone P, Dionigi R, Bellinzona G, Imberti R, Stocchetti N (1998) Effects of cerebral perfusion pressure on brain tissue PO2 in patients with severe head injury. Acta Neurochir 71 (Suppl):111113 11. Lang EW, Czosnyka M, Mehdorn M (2003) Tissue oxygen reactivity and cerebral autoregulation after severe traumatic brain injury. Crit Care Med 31:267271 12. Shaffner DH, Eleff SM, Brambrink AM, Sugimoto H, Izuta M, Koehler RC, Traystman J (1999) Effect of arrest time and cerebral perfusion pressure during cardiopulmonary resuscitation on cerebral blood flow, metabolism, adenosine triphosphate recovery, and pH in dogs. Crit Care Med 27:13351342